An intelligent prediction method for aerodynamic performance of an automobile

Abstract

The application provides a kind of intelligent prediction method for automobile aerodynamic performance, comprising: obtaining vehicle shape structure parameterization data, flow field simulation data, flow field test data, and building a data system;Filtering and denoising the data in the database, extracting the characteristic data strongly related to aerodynamic performance, and performing data analysis and fusion;Realize intelligent prediction of aerodynamic performance parameters;Update the automobile aerodynamic prediction function to realize the prediction of automobile aerodynamic performance.Through the establishment of automobile research and development field aerodynamic digital development application scene, it has wide industry application value in reducing development cost, improving development efficiency and quality, etc.The technology shows that, in the whole development process, by using machine learning and digital means, it can provide efficient, low-cost and high-quality digital solutions for automobile enterprises, and has certain industry demonstration effect.

Description

Technical Field

[0001] This invention belongs to the field of automotive R&D technology, and in particular relates to an intelligent prediction method for automotive aerodynamic performance. Background Technology

[0002] Under the dual-carbon development goals, automotive aerodynamics development has become a crucial technology for automakers to meet the new challenges of energy conservation and emission reduction. By building an aerodynamically driven digital development platform, aerodynamic technology can be integrated with automotive styling design, revealing the mechanisms and laws governing the influence of vehicle body flow fields on automotive aerodynamics. This enables a shift from a human-centric R&D model to a data-driven one, resulting in a systematic digital aerodynamic solution.

[0003] The traditional development process for automotive aerodynamics is mainly based on manual iteration. During the vehicle design phase, the aerodynamics team, styling team, and engineering design team coordinate and iterate the design. This process is characterized by long design cycles and high design costs. The optimization process occurs in the middle and late stages of vehicle development, lacking early development methods and tools. Without effective data support, there are no methods or means to quickly and efficiently predict aerodynamic performance, resulting in high development difficulty and affecting the achievement of goals.

[0004] Currently, aerodynamic performance development relies primarily on engineers' experience, analyzing and optimizing based on limited flow field information. While it possesses some wind tunnel testing capabilities, it lacks the ability to mine and intelligently process flow field information. Data accumulation is limited and slow, restricting development capabilities and making it difficult to improve the quality of development projects, thus hindering innovation and expansion in the aerodynamics field.

[0005] With the development of digital technology, automotive R&D is facing significant technological changes. Utilizing machine learning and digital methods to achieve vehicle aerodynamics development is a key common problem facing aerodynamics. Currently, there is no such digital aerodynamics development platform in China, but the industry has an urgent need for digital transformation in aerodynamics R&D. Summary of the Invention

[0006] In view of this, the present invention aims to propose an intelligent prediction method for automotive aerodynamic performance, so as to quantitatively evaluate the impact of automotive structural parameters on aerodynamic performance, which is of great significance for guiding vehicle design and improving the efficiency of aerodynamic performance development.

[0007] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0008] A method for intelligent prediction of automotive aerodynamic performance includes the following steps:

[0009] S1. Obtain parameterized data of the vehicle's external structure, flow field simulation data, and flow field test data;

[0010] S2. Based on the data obtained in step S1, build a data system;

[0011] S3. The integrated Kalman filter method is used to filter and denoise the data in the database, extract feature data that is strongly related to aerodynamic performance, and perform data analysis and fusion.

[0012] S4. Using nonlinear fitting or artificial neural network methods, establish the functional relationship between the aerodynamic performance objective function and various variables such as vehicle structural parameters, flow field pressure, and flow field velocity to achieve intelligent prediction of aerodynamic performance parameters.

[0013] S5. The difference between the predicted results and the measured results is analyzed using the root mean square error. Based on the analysis results, the automotive aerodynamic prediction model is calibrated and optimized, thereby updating the automotive aerodynamic prediction function and realizing the prediction of automotive aerodynamic performance.

[0014] Furthermore, in step S1, parameterized data of the vehicle's external structure are obtained through an automatic recognition algorithm, vehicle flow field simulation data are obtained through a digital wind tunnel method, and vehicle flow field test data are obtained through an aeroacoustic wind tunnel method.

[0015] Furthermore, the external structural parameters in step S1 include angle data, height data, and angle and distance data between various parts of the vehicle.

[0016] Furthermore, in step S2, the database system obtains vehicle aerodynamic performance parameters through wind tunnel testing, road testing, and CFD simulation;

[0017] Vehicle aerodynamic performance parameters include aerodynamic forces, wind speed around the vehicle, and pressure distribution.

[0018] Furthermore, in step S3, Gaussian regression is used to fuse the data obtained from simulation and experiment.

[0019] Furthermore, the functional relationship formula in step S4 is as follows:

[0020] Cd=k*f(X n ,P n V n ), where C d The drag coefficient is Xn = (x1, x2, ... xn). n ) represents the geometric structural parameters for automobile styling; P = (p1, p2, ... p n ) represents the pressure within the flow field; V = (v1, v2, ... v n ) represents the velocity within the flow field.

[0021] Furthermore, in step S4, during the intelligent prediction of aerodynamic performance parameters, an adaptive spatial transformation is used to scale the known parameters.

[0022] Compared with existing technologies, the intelligent prediction method for automotive aerodynamic performance described in this invention has the following advantages:

[0023] (1) The intelligent prediction method for automotive aerodynamic performance described in this invention has broad industry application value in terms of reducing development costs, improving development efficiency and quality by building a digital development application scenario for aerodynamics in the automotive R&D field.

[0024] (2) The intelligent prediction method for automotive aerodynamic performance described in this invention demonstrates that by using machine learning and digitalization throughout the entire development process, automotive companies can be provided with efficient, low-cost and high-quality digital solutions, which has a certain industry demonstration effect.

[0025] (3) The intelligent prediction method for automotive aerodynamic performance described in this invention changes the traditional human-based performance development model to an intelligent development model. The reuse of models and data and the accumulation of knowledge will greatly improve the development efficiency and quality of aerodynamic performance. This invention has good innovative value. Attached Figure Description

[0026] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0027] Figure 1 This is a schematic diagram of the vehicle aerodynamic performance prediction function according to an embodiment of the present invention;

[0028] Figure 2 This is a schematic diagram of the parameters of the automobile structure described in an embodiment of the present invention. Figure 1 ;

[0029] Figure 3 This is a schematic diagram of the parameters of the automobile structure described in an embodiment of the present invention. Figure 2 . Detailed Implementation

[0030] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0031] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0032] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0033] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0034] A method for intelligent prediction of automotive aerodynamic performance includes the following steps:

[0035] Step 1: Data Acquisition and Collection. Using technologies such as automatic recognition algorithms, digital wind tunnels, and aeroacoustic wind tunnels, parameterized data (x1, x2, ... xn) of the vehicle's external structure, flow field simulation data (pressure (p1, p2, ... pn), velocity (v1, v2, ... vn), drag coefficient Cd, lift coefficient Cl, lateral force coefficient Cs, etc.), and flow field test data (pressure (p1, p2, ... pn), velocity (v1, v2, ... vn), drag coefficient Cd, lift coefficient Cl, lateral force coefficient Cs, etc.) are obtained. The automatic recognition algorithm can automatically identify and measure geometric data of different formats to obtain external structural parameters, including angle data, height data, and angle and distance data between vehicle components. A digital model of the vehicle and the wind tunnel laboratory is established in the digital wind tunnel, and computational fluid dynamics algorithms are used to obtain flow field simulation data around the vehicle, including pressure, velocity, and aerodynamic coefficients. The aeroacoustic wind tunnel uses wind tunnel testing technology to obtain flow field test data, including pressure, velocity, and aerodynamic coefficients.

[0036] A digital wind tunnel includes a simplified geometric model of the wind tunnel chamber, a road surface simulation equipment model, and a model of the vehicle under test.

[0037] Step 2: Build a database system using the collected parametric data of vehicle shape structure, flow field simulation data, and experimental data;

[0038] The database system contains vehicle aerodynamic performance parameters obtained from wind tunnel tests, road tests, CFD simulations, etc., including aerodynamic forces, wind speed around the vehicle, pressure distribution, etc.

[0039] Step 3: Use the integrated Kalman filter method to filter and denoise the data in the database, extract feature data that is strongly correlated with aerodynamic performance, and perform data analysis and fusion.

[0040] Data fusion employs Gaussian regression to fuse data obtained from simulations and experiments;

[0041] Step 4: Using nonlinear fitting or artificial neural networks, establish the functional relationship between the aerodynamic performance objective function and variables such as vehicle structural parameters, flow field pressure, and flow field velocity, Cd=k*f(Xn,Pn,Vn), to achieve intelligent prediction of aerodynamic performance parameters;

[0042] Where Cd is the drag coefficient; Xn=(x1,x2,...xn) are the geometric structural parameters of the car styling; P=(p1,p2,...pn) is the pressure in the flow field; V=(v1,v2,...vn) is the velocity in the flow field;

[0043] In the process of intelligent prediction of aerodynamic performance parameters, adaptive spatial transformation is used to scale the known parameters;

[0044] Step 5: Analyze the difference between the predicted results and the measured results using the root mean square error. Based on the analysis results, calibrate and optimize the vehicle aerodynamic prediction model to update the vehicle aerodynamic prediction function and realize the prediction of vehicle aerodynamic performance.

[0045] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent prediction of automotive aerodynamic performance, characterized in that... Includes the following steps: S1. Obtain parameterized data of the vehicle's external structure, flow field simulation data, and flow field test data; S2. Based on the data obtained in step S1, build a data system; S3. The integrated Kalman filter method is used to filter and denoise the data in the database, extract feature data that is strongly related to aerodynamic performance, and perform data analysis and fusion. S4. Using nonlinear fitting or artificial neural network methods, establish the functional relationship between the aerodynamic performance objective function and various variables such as vehicle structural parameters, flow field pressure, and flow field velocity to achieve intelligent prediction of aerodynamic performance parameters. S5. The difference between the predicted results and the measured results is analyzed using the root mean square error. Based on the analysis results, the automotive aerodynamic prediction model is calibrated and optimized, thereby updating the automotive aerodynamic prediction function and realizing the prediction of automotive aerodynamic performance. The external structural parameters in step S1 include angle data, height data, and angle and distance data between various parts of the vehicle; In step S3, the data fusion uses Gaussian regression to fuse the data obtained from simulation and experiment. The functional relationship formula in step S4 is as follows: Cd=k*f(X n ,P n V n ), where C d X is the drag coefficient; n =(x1,x2,...x n ) represents the geometric structural parameters of the vehicle's styling; P n =(p1,p2,...p n V represents the pressure within the flow field; n =(v1,v2,...v n () represents the velocity within the flow field; In step S4, during the intelligent prediction of aerodynamic performance parameters, an adaptive spatial transformation is used to scale the known parameters.

2. The intelligent prediction method for automobile aerodynamic performance according to claim 1, characterized in that, In step S1, the vehicle's external structure parameter data is obtained through an automatic recognition algorithm, the vehicle's flow field simulation data is obtained through a digital wind tunnel method, and the vehicle's flow field test data is obtained through an aeroacoustic wind tunnel method.

3. The intelligent prediction method for automobile aerodynamic performance according to claim 1, characterized in that, In step S2, the database system obtains vehicle aerodynamic performance parameters through wind tunnel testing, road testing, and CFD simulation. Vehicle aerodynamic performance parameters include aerodynamic forces, wind speed around the vehicle, and pressure distribution.

Images